Semantic actions/attributes and LR parsers

cowboy@badlands.cv.lexington.ibm.com (Rick Nelson)
Tue, 21 Mar 1995 15:27:25 GMT

          From comp.compilers
Related articles
Semantic actions/attributes and LR parsers cowboy@badlands.cv.lexington.ibm.com (1995-03-21)
| List of all articles for this month |
Newsgroups: comp.compilers
From: cowboy@badlands.cv.lexington.ibm.com (Rick Nelson)
Keywords: parse, attribute
Organization: IBM Austin
Date: Tue, 21 Mar 1995 15:27:25 GMT

Hi,


The approach for implimenting semantic actions (introducing epsilon
productions) in LR parsers can cause grammar conflicts. Attribute
grammars also are problematic.


This fact is well known, but how often does it really come up? How
much of a pain is it to re-work the grammar (or do you switch to an
LL based parser?)


I'm looking for examples, real or contrived (fractions of grammars
OK) where this happens - and the hack(s) required to bypass it.


Please email if possible, I'd like to combine/summarize (will munge
grammars or elide your example if requested).


Rick Nelson
--

Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.